|Reza Baba ghasabha|
Design and Implementation of Vision Based Robust Nonlinear Controller on KNTU CDRPM
Cable driven parallel robots (CDPRs) remedy some shortcomings of the conventional serial and parallel robots. Using cables instead of rigid links in the robot structure has some positive features such as large workspace, high speed manipulation and high payload to robot weight ratio. However, using cables introduces new challenges in the study of cable-driven parallel robots. Cables are able to apply only tensile forces and due to the inherent ﬂexibility of the cables, measuring cable length is not reliable in applications with high accuracy. Furthermore, due to the axial vibrations in cables, the moving platform may experience unwanted vibrations. Moreover, it shall be noted that in practice the kinematic and dynamic models of CDPRs possess structured and parametric uncertainties and precise knowledge of the models is unavailable. For these reasons, control of the cable driven parallel robots is more challenging than that of the conventional robots. In this thesis, new approaches is proposed in order to overcome these challenges and improve the performance of the CDPRs in tracking the desired trajectories. Due to the inaccuracy of the cable length measurement, a vision-based pose measurement is designed and implemented as a suitable and economical solution. An appropriate marker is designed for fast and accurate tracking and the pose of the end-eﬀector is measured by extracting marker features in real-time. Then, an adaptive controller is designed in task space coordinates and adaptation is performed on both the dynamic and kinematic parameters. It is shown that the performance of the proposed adaptive controller is signiﬁcantly improved by the correction of the internal forces. Then, In order to reduce the eﬀect of both structured and parametric uncertainties and external disturbances, an adaptive robust controller is designed based on the adaptation of the uncertainties upper bound. The main feature of this approach is that the result is not in debt of ﬁnding a linear regression form for kinematic and dynamic models, and furthermore, it does not require a priori knowledge on the uncertainties upper bound. Then, in order to damp the longitudinal vibration of cables, a composite controller is proposed. By using the result of singular perturbation approach, the stability of the overall closed-loop system is analyzed through Lyapunov second method. Finally, the eﬀectiveness of the proposed controllers is examined through some experiments on a planar cable driven parallel robot with four actuated cable driven limbs and it is shown that the proposed controllers is able to provide suitable tracking performance in practice.
|2014||M.Sc.||Parallel and Cable Robotics|